MINDSETOngoing practice

Failure as Investment

Treat mistakes as the cost of innovation, not as evidence of incompetence

Problem it solves

build psychological safety in risk-averse teams"

Best for

["Organizations doing novel or creative work where experimentation is essential","Leaders who need to build psychological safety in risk-averse teams","Teams paralyzed by perfectionism or fear of judgment","R&D environments where dead ends are part of the discovery process"]

Not ideal for

["Contexts where failures have irreversible safety consequences (surgery, aviation)","Compliance-heavy environments where error tolerance is genuinely near zero","Teams that lack basic competence and use 'failure is okay' as an excuse for sloppiness"]

Overview

Why this framework exists

Catmull argues that failure is not a necessary evil but an inevitable and valuable consequence of doing anything new. The deep-seated shame around failure, traced back to childhood schooling, causes people to avoid risk, which leads to derivative, uninspired work. The real danger is not failure itself but the organizational response to failure: when errors trigger blame-seeking rather than learning, people shut down and stop experimenting.

Andrew Stanton's mantra at Pixar was 'fail early and fail fast' and 'be wrong as fast as you can.' The subtle point is not to accept failure with resignation but to treat it as aggressive, rapid learning. Moving forward, even in the wrong direction, gives the team momentum and reveals information that standing still never would. Indecision, driven by the desire to avoid failure, is the far worse mistake.

Catmull frames production meltdowns as investments in R&D rather than catastrophes. When Pixar had to replace directors on Cars 2 and Monsters University, and shut down another project entirely, those losses taught lessons that improved future work. The key leadership behavior is being open about your own mistakes, which signals to others that it is safe to take risks.

Core principles

6 total
  1. Mistakes are not evil; they are an inevitable consequence of doing something new and should be seen as valuable
  2. The desire to avoid failure is more dangerous than failure itself because it kills experimentation
  3. Being open about mistakes as a leader makes it safe for others to take risks
  4. Moving forward in the wrong direction yields more information than standing still
  5. The cost of preventing all errors is far greater than the cost of fixing them
  6. Meltdowns are investments in organizational learning, not evidence of systemic dysfunction

Steps

5 steps
  1. Reframe failure linguistically and culturally
    Stop using failure as a synonym for incompetence. Explicitly tell your team that mistakes are expected when doing new work. Use phrases like 'exploring the neighborhood' to describe dead ends that yield useful information.
  2. Model vulnerability from the top
    As a leader, publicly discuss your own mistakes and your role in organizational failures. When people see that leadership treats its own errors as learning opportunities rather than career threats, they become willing to take risks.
  3. Lower the cost of individual experiments
    Like Andrew Stanton's analogy of learning to ride a bike with training wheels and pads, structure work so that early experiments are cheap and fast. Show rough work early, test ideas in low-stakes settings, and iterate before committing large resources.
  4. Eliminate the blame reflex
    When an error is discovered, the first question should be 'What can we learn?' not 'Whose fault is it?' If people shut down and turn inward after mistakes, your culture vilifies failure regardless of what your mission statement says.
  5. Budget for failure explicitly
    Treat production meltdowns and abandoned projects as line items similar to R&D spending. When failure has an acknowledged budget, it loses its stigma and teams can make rational decisions about when to pivot or kill a project.

Examples

1 cases
Monsters, Inc. evolution through dead ends

Pete Docter's original concept for Monsters, Inc. was about a thirty-year-old accountant haunted by monsters representing childhood fears. Over years of development, the protagonist changed from a man to a six-year-old girl named Mary, then to a boy, then back to a girl, then to a bossy seven-year-old named Boo, and finally to a preverbal toddler. The sidekick character Mike was not added until over a year into development. Each 'wrong' version yielded insights that led closer to the final story.

OutcomeThe finished film bore no resemblance to the original pitch but was a massive critical and commercial success. The team never treated failed approaches as personal defeats; each dead end narrowed the search toward the right story.

Common mistakes

2 traps
Interpreting 'fail fast' as 'accept mediocrity'
Many people hear 'failure is okay' and conclude that low standards are acceptable. The real meaning is to iterate aggressively and learn quickly, not to stop caring about quality. The goal is excellence through rapid experimentation, not resignation to poor outcomes.
Searching for scapegoats after failures
When organizations investigate failures primarily to assign blame, people learn to avoid anything novel. The Golden Fleece Award phenomenon Catmull describes shows how public shaming of failed experiments distorts future risk-taking across an entire institution.

Origin story

How this framework came to be

Catmull developed this framework after observing how the Golden Fleece Awards in the 1970s created a chilling effect on scientific research by publicly shaming projects that appeared wasteful. He realized that vilifying failure distorted how researchers chose projects. At Pixar, the principle crystallized during the string of production crises following Toy Story 3, when several films required director replacements or were shut down. Rather than treating these as disasters, Catmull reframed them as necessary investments, and the studio emerged stronger from each one.

Source

Traced to primary
Source · BOOK
Creativity, Inc.
Ed Catmull & Amy Wallace · 2014
Open source →

Related frameworks

Browse all Mindset →